JP2018127882A - Method of optimizing track - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a method of maneuvering a railroad maintenance machine capable of optimizing the track position and running on the track after the drive test by a track measuring vehicle which can run on the track.SOLUTION: For the purpose of providing a suitable measuring behavior, a test drive is carried out by a track measuring vehicle provided with an inertia navigation device which is capable of running on the track in the first place. A wheel of the track measuring vehicle is pressed against a rail to measure the length of a curve by an odometer so that angular position of the inertia navigation device with respect to the cartesian coordinate systems x, y, z is measured and stored. After the test drive, the coordinate point of the track of the inertia navigation device in the coordinate space is calculated based on the stored angular position by interpolation calculation. Based on the coordinate point of the track that is optimized in such a manner, the position graph is created by setting z-element to zero and the longitudinal section is generated by setting x-element to zero, followed by drawing a single curve calculated by mathematical associative function over the coordinate points.SELECTED DRAWING: Figure 7

Description

  The present invention optimizes the track position, measures at least two axles supported by one machine frame, one navigation device, one piece of equipment for pushing the wheels against the rail, and the length of the curve of the rail. The present invention relates to a method for operating a railway maintenance machine capable of traveling on a track after measurement traveling by a track measuring vehicle capable of traveling on a track having an odometer.

  Many railroad tracks are configured as ballast tracks. In this case, sleepers are present in the ballast. Irregular sleeper settling and displacement of the lateral position of the track are caused by the acting forces of the wheels of the train running on the track. Vertical height errors, cant errors (within the curve), and directional position errors are caused by ballast subsidence. When the predetermined riding comfort limit values or safety limit values of these shape variables are exceeded, maintenance work is performed.

  Today, in many cases, removal and error of these shape trajectory errors are performed by railway maintenance machines. The railroad maintenance machine is equipped with a so-called inspection measuring device and inspection recording device so that the track can be used again without any problem after the work for improving the track shape. Test tolerances are specified for track quality after track maintenance or improved by such methods. The verification tolerance indicates a minimum requirement for the quality of the shape improvement made. The minimum requirement is inspected by the inspection and measurement apparatus and the inspection recording apparatus. Important variables to be corrected and recorded at this time are deviation, vertical height of the trajectory, trajectory direction or lateral position, and lateral trajectory or cant of the trajectory. Railway maintenance machines, such as tamping machines, correct the track shape that has deteriorated due to the train load. For this reason, the trajectory is adjusted to the target position by the lifting / lowering adjustment device controlled by the electrohydraulic method.

  The smaller the residual error after the maintenance work of the track maintenance machine, the smaller the interaction force between the vehicles, and the track shape deteriorates more slowly during vehicle operation, and the invariance of the track position becomes larger. It is therefore desirable to make the trajectory shape as close as possible to its target position. This is because considerable costs and labor associated with it can be saved. The faster the train runs, the smaller the tolerance value. Correction of errors (generally 3 m to 70 m wavelength error) is important for a traveling train. The wavelength error to be considered is adjusted according to the section speed.

  Various trajectory correction methods have been developed to correct trajectory errors. On the one hand, there is a relative reference construction that smoothes the track position exclusively, and on the other hand, there is an absolute reference construction. In this absolute reference construction, the track position is corrected according to a preset target shape. The target shape of the railway track is provided as a track position design drawing and can be used while confirming the behavior of the measurement system to calculate a system error after being input to the control computer of the track maintenance machine. If the absolute correction value for the front end of the machine measuring device is known, the front end of the machine measuring device is guided on its target curve and the rear end is guided on the already corrected trajectory. The correction process is performed at the work position. The position of the tamping machine in the longitudinal direction of the track is usually measured by an odometer.

  If the target shape is not known, using known methods, the trajectory position is often measured by a directional string mechanism and a vertical height string mechanism prior to work. In this case, the actual lateral tilt is measured by the pendulum. The pendulum depends on the acceleration generated during the measurement run. In particular, the centrifugal acceleration during measurement running on a curve causes a large error. As a result, the measurement speed is generally limited to a range of about 5 km / h. The measurement of the trajectory position by the string mechanism is affected by the transfer function. That is, the measured signal differs from the actual trajectory error with respect to its waveform, strength and phase position. The measured value is proportional to the curvature of the trajectory. Similarly, measurement of the orbital height position by the string mechanism provides a signal proportional to the curvature affected by the transfer function. The gradient change point (the transition from one slope to the other slope) cannot be calculated from the measurement data. This is because the transition portion of the slope that is recognized as the curvature difference as well as the curvature error of the orbital position cannot be identified. Furthermore, the actual slope cannot be estimated from the curvature measurement of the height (but the tracker at a given slope cannot derive the actual slope from the curvature measurement because the measurement system Because all of this is within the slope and only relative measurements are performed).

  In order to optimize the trajectory position according to known conventional methods, the measured curvatures are smoothed into one target curvature curve. The difference between the measured curvature graph and the smoothed “target curvature graph” is then generated. An approximated inverse transfer function is generated by a digital filter (see DE 10337976). This inverse transfer function is then applied to the difference between the measured curvature graph and the smoothed “target position curvature graph”. Thereby, an orbital error is approximately obtained. When the error to be corrected exceeds the maximum allowable correction value, the “target position curvature graph” needs to be appropriately changed by a complicated method. Similarly, the fixed position or fixed point needs to be changed appropriately. A fixed position such as a bridge is not allowed the necessary lateral movement. In the case of the height in the vertical axis direction, a different processing step occurs, unlike the direction. Because it can be raised exclusively, it cannot be lowered. The transition between the rise and fall resulting from the optimization needs to be positively defined. Furthermore, the up correction value is raised so that negative (down) no longer occurs. Therefore, the calculation is difficult and inaccurate. This is because the starting point is the generation of curvature. The error almost coincides with the second-order integral of the curvature difference, but this second-order integral is similarly affected by the transfer function.

  Conventional techniques are string measuring mechanisms and pendulums or clinometers for measuring vertical height, direction and lateral tilt. The conventional technique is a plurality of odometers for measuring the position of the measurement mechanism on the track. The prior art is also a satellite navigation system (such as GPS, Navstar or Galileo). Further, the conventional technology is an inertial navigation system (INS) or inertial navigation device including a central sensor device having three acceleration sensors and a rotational speed sensor. By integrating the acceleration and rotational speed measured by the IMU (Inertial Measurement Device), the three-dimensional movement of the vehicle is continuously measured in the INS, and from this measurement, each geographical position is continuously Measured. The INS system operates at a data rate of about 100-1000 Hz, high accuracy, and slight drift (0.01 ° -less than 0.05 ° / hour). If the INS system does not move, the INS system will automatically calibrate while it is stopped. The main advantage of an INS is that it can be operated without reference. The acceleration may be measured by an acceleration sensor fixed to the vehicle (“in a strap-down manner”). Of course, in principle, only one IMU may be used. In this case, the absolute roll angle needs to be measured by an independent clinometer. The advantage of these measuring systems is the roll angle that can be measured independent of centrifugal acceleration, the system's = 1 transfer function valid within a wide range. That is, the actual track of the vehicle is measured in the coordinate space without distortion of the waveform, strength or phase position of the track error. A three-dimensional coordinate line is acquired from this three-dimensional trajectory of the vehicle in the coordinate space and measurement at equal intervals by an odometer. A position graph of the trajectory is obtained by projection onto the xy plane, and a vertical sectional view is obtained by projection onto the yz plane. In addition, satellite navigation data (eg by GPS) can be recorded. The conventional technology is an INS system based on the so-called “north”. The INS system provides absolute angular deviation of roll angle, yaw angle and pitch angle for systems oriented north. In this case, the y unit vector is directed so that the x unit vector indicates north, the z unit vector indicates the direction of gravity, and an orthonormal system is established. The angular deviation represents a unit vector indicating the direction of the measurement vehicle in which the INS system exists.

  In the case of the known method for optimizing the trajectory position from the measurement run and correcting the trajectory error, the measured value affected by the transfer function, the inverse transfer function to be expressed approximately exclusively for the string measuring mechanism, the action from the outside Disadvantages are the adverse effects of the clinometer due to acceleration, and boundary conditions such as fixed positions, fixed points or maximum permissible correction values during optimization calculations that are difficult to satisfy. The small measuring speed allowed during the measurement run is also a drawback. Another disadvantage is that the string measuring mechanism extends locally 10-20 m in the machine direction. Therefore, in many cases, the track maintenance machine includes a support portion. A measuring device is stored under the support. As a result, the cost and length of the machine increases. If additional work equipment such as a sweeper, ballast hopper or dynamic orbital stabilizer is attached to the support, the use of a string measuring mechanism is not possible. The known method for optimizing the trajectory position has the disadvantage that the absolute inclination of the trajectory and the gradient change point cannot be determined from the measurement data. According to the known method, since the position of the gradient change point is not known, the trajectory position correction including an error is executed within the range. Another disadvantage is that, in the case of the known method, the trajectory to be actually corrected due to the non-regressive digital filter used and the installation spacing which is approximately 13 times the short chord interval required for the filter. (In general, a is in the range of 4-7 m, so that the front measurement length and the rear measurement length are 50-90 m). In the case of the small measuring speed that the known method exhibits, therefore, a corresponding time is required. Furthermore, the known string method must be regularly controlled (eg due to wear of the measuring wheel, inaccuracies in the transducer, inaccuracies in the support from which the string extends, string vibrations, etc. ) Has the disadvantage of offset. Recalibration of the string measuring mechanism is cumbersome and expensive.

German patent No. 10337976

  Therefore, the subject of the present invention is the influence of externally acting acceleration, the limited measurement speed, the influence of the transfer function of the string measurement mechanism, the offset to be calibrated of the string measurement mechanism and the boundary conditions that are difficult to satisfy. In order to avoid the disadvantages of the above known systems and the installation of additional supports, using an inertial navigation device (INS) or inertial measurement device (IMU) and odometer, fixed positions, fixed points and maximum permissible In consideration of boundary conditions such as the track position correction value, the method for optimizing the track position is to be improved according to the measurement travel.

  This problem is solved by the features of claims 1 to 10. Other preferred configurations of the invention are described in the dependent claims.

  For this purpose, the INS system is installed on one measuring vehicle having two axles that are configured to be rotatable relative to each other. Furthermore, the measurement vehicle can be configured such that these axle rotation devices include a track width measurement unit. Thus, the geometric position of the two rails can be measured during the measuring process. Wheel spacing points can be measured by knowing exactly the track width. As a result, the Kant measurement accuracy is improved. An odometer for measuring the travel distance of the measuring vehicle traveling on the track is coupled to the measuring vehicle. The measurement vehicle is brought into contact with the side surface of the rail during measurement, or is pressed on both sides by the track width measurement unit. Therefore, the INS measures the tangent in the track direction, the vertical axis inclination and the horizontal inclination (cant) of the measurement vehicle on the track. INS measurement data for the corresponding point, for example (generally 0.25, 0.5 or 1 m, and almost continuous recording is possible due to the high measurement speed of the INS). Is memorized. In addition to the INS data, the length of the accurately advanced curve (or “start-up kilometer”) is also stored for each measurement point.

  The present invention provides a target trajectory position optimized in height and direction, and provides a correction value suitable for the optimized target trajectory position according to the length of the curve measured by the odometer. Another advantage other than the higher measurement speed is the robustness of the INS or IMU against general disturbance acceleration. The ballast track is tamped at 35 Hz and the track maintenance machine is driven by a diesel engine and therefore has disturbing vibration and acceleration levels. Furthermore, the method according to the invention provides the advantage that the position of the slope change point and the absolute slope can be measured accurately, so that a quantitatively more accurate correction value can be set for the railway maintenance machine. The higher achieved track position accuracy extends the durability life of the track shape and significantly reduces maintenance costs. Another advantage is that the front measurement length and the rear measurement length are almost constrained to the length of one machine (generally 10-20 m). In addition, an accurate curvature graph and an inclination graph in the vertical axis direction corresponding to a conventional graph of a railway track can be created by the method of the present invention. Therefore, in the following discussion, the target data will be described. Furthermore, the measurement data provides an excellent criterion for the measurement work of railway tracks to determine the target shape of the still unknown track shape of the track. Since the target deviation and the target height in the vertical axis direction are mathematically calculated from the target position graph and the target slope graph, the method of the present invention does not beneficially cause an offset of the measurement system. Therefore, a troublesome calibration method is omitted.

  After the measurement run is completed, the trajectory of the measurement vehicle in the coordinate space is the absolute angle difference with respect to the coordinate system based on the north of the INS values (roll angle, yaw angle and pitch angle) recorded for each measurement point. Calculated from As a start condition, a start vector is preferably assigned to the start direction of the measurement vehicle relative to the start of the measurement run. As described below, the integration of the trajectory (calculation of three-dimensional coordinates of the measured travel position) is executed.

  The position graph f (xy) is obtained by setting z = 0 in the three-dimensional coordinates. A longitudinal sectional view g (y, z) is created by setting x = 0. This longitudinal sectional view shows the height transition, and also shows the slope change point (the transition from one slope to the other slope). A reference length is defined depending on the track grade. The position graph is divided into a plurality of sections according to a reference length set in advance according to the line type. Smoothing is performed according to a method of adding least squares of a plurality of deviations using, for example, a plurality of three-dimensional spline curves. These spline curves are combined with each other into a single smoothed target position curve. Subsequently, the plurality of correction values are cut off by the maximum allowable correction value. The plurality of fixed points are combined by one mathematical function that passes through the plurality of transition points and the plurality of fixed points in succession. The immediately preceding target position curve is transformed into a target position curve newly changed by superposition. A slight change is maintained without any influence on the running of the vehicle. The same processing is performed for the fixed position. As a result, the target position curve to be changed and used is finally obtained. The difference with respect to the target curve in this direction is a rectification correction value transmitted to the track maintenance machine. This target curve needs to be further converted into a string measurement mechanism value (deviation relative to that direction) so that the track holding machine can process the target curve in that direction using that string mechanism. This conversion is performed computationally. In the conversion, the string is continuously included in the target curve in the direction. The target deviation calculated in this way is transmitted to the control unit of the track maintenance machine. Therefore, the track maintenance machine can perform trajectory correction in the direction.

  For height (set to x = 0), a height transition (g (y, z)) is created at the absolute height position according to the present invention. The transition from one slope to the other indicates the slope change point. This slope change point can be easily calculated in the filtered data of the actual height transition by a digital filter that allows only long wavelength changes (eg, the second derivative of the filtered height transition is , Make it possible to calculate the slope change point clearly). When the gradient change point is calculated, a plurality of regression functions (for example, a plurality of regression lines) are calculated between the plurality of gradient change points. These individual regression curves are combined by a higher order interpolation curve, for example a three-dimensional spline curve (generally having a predetermined inner radius length, for example 6 m). The result is a continuous mathematical function as the target height transition. Subsequently, a difference between the measured actual height transition and the target height transition is generated. This difference corresponds to a correction value to be executed and is raised so that a negative value does not occur and a purely positive rise value is obtained. The positive bottom-up value needs to be further converted into a string measuring mechanism value (vertical height) so that the track maintenance machine can process the target height transition with its string mechanism. This conversion is performed computationally. In the conversion, the string is continuously included in the target curve at the height position. The target height in the vertical axis direction calculated in this way and the raised value are transmitted to the controller of the track maintenance machine. Therefore, the track maintenance machine can perform trajectory correction at the height in the vertical axis direction.

  When the operation is completed, a corrected trajectory position corresponding to the transition of the target position curve and the target height transition is obtained while maintaining boundary conditions such as a fixed position, a fixed point, or a maximum correction value.

  The system described here can of course be used with the advantage that there is no spatially wide measuring system and therefore no support is required as a unique measuring system for creating inspection records. Good. Thus, in connection with this method, the advantage is that the measurement can be recorded at a speed reaching 80 km / h. As a result, the time taken for the measurement run is significantly reduced compared to existing systems which are generally executed exclusively at about 5 km / h. The curvature is obtained when the position graph is mathematically differentiated twice according to the length of the curve. The curvature graph is a notation commonly used on railway tracks. Another advantage is that a predetermined change in position of the track (eg, a turn in the station premises) can be directly superimposed on the target position graph. At this time, the track maintenance machine can easily change the target position graph after receiving the control data obtained from the target position graph. Such a preset in the curvature graph is not possible. Furthermore, according to the method of the present invention, the position of the slope change point and the absolute slope can be calculated accurately, so that a qualitatively more accurate correction value can be preset for the railway maintenance machine. Is obtained.

  The subject matter of the present invention is schematically illustrated in the drawings.

It is the top view and elevation which show the trajectory measuring apparatus of this invention which has an inertial navigation apparatus. It is a principle figure of the highly accurate orbit correction method on the basis of a string. It is a figure of the three-dimensional orbit coordinates on the basis of the north. FIG. 3 is a diagram of three-dimensional trajectory coordinates representing a position graph and an inclination graph in the vertical axis direction. It is a figure which calculates | requires the target track | orbit position optimized by the three-dimensional interpolation spline curve. It is a detailed view of a smooth curve between a plurality of interpolation spline curves. It is the schematic which calculates | requires a target track position in consideration of only one fixed point. It is the schematic which calculates | requires a target track position in consideration of a fixed position. It is a vertical sectional view in the vertical axis direction having the interpolated target height position, gradient change point, and height correction value.

FIG. 1 shows a track measuring vehicle A having an inertial navigation device 5 arranged on a measuring vehicle frame 8 which is coupled according to the invention. The measurement vehicle A travels on the rail 12 by wheels. A plurality of wheels 1 on one side are fixedly coupled to each other 2, 3. Two opposite wheel side surfaces are supported by a shaft 4 so as to be able to twist and rotate. Therefore, the track measuring vehicle A occupies a predetermined position for each track error. The shaft 4 is configured to be slidable in the axial direction. The sliding is performed by the pneumatic cylinders 6 and 14 so that both the left and right wheel pairs 2 and 3 travel along the rail side edge. The actual gauge SPW is measured by the distance sensor 7. The wheel interval RAW is calculated as follows.
RAP = SPW + b _k
b _k ... rail head width

  Vertical and horizontal stresses are applied to the wheel pairs 1, 2 and 1, 3 by the two tilt cylinders 10. This ensures the proximity of the wheel 1 to the rail 12 and avoids derailment. A vertical cylinder is used to lift the measuring vehicle in the locked position and place it on the rail in order to start or end work. Reference numeral 11 denotes a friction bearing. Within this friction bearing, the torsion shaft 4 can be twisted and slid laterally. Reference numeral 21 denotes a machine frame of the track maintenance machine. An odometer (odometer) 24 that measures the length of the curve of the track is attached to the measurement vehicle A.

FIG. 2 schematically shows the operating principle of the high-precision trajectory correction system B. This high-precision trajectory correction mechanism B comprises a string of length L having string sections a and b. The track maintenance machine having the correction mechanism based on the string operates in the AR direction. Reference numeral 21 indicates a trajectory position to be corrected, and _RS indicates an optimized trajectory target position. In the illustrated graph, the trajectory behind the string has already been corrected. In this case, the rear point of the string exists on the corrected trajectory. However, the front point exists on the target trajectory position 21 ahead. That is, for the method, it is necessary to calculate the correction value v first. The front chord end needs to be guided on the trajectory target position _RS . For example, this can be accomplished by the string being guided using a forward lateral displacement device. The string is displaced exactly by the difference v during each operating step. Another condition for the function without the drawbacks of the method is a preset of the target deviation f _S. Therefore, in order to obtain the trajectory target position f _S ,
RW ₌ f s -f _j
RW ... A reference value is required as a quasi-value.

That is, for a function without a defect, the target deviation f _S and the correction value v need to be calculated by an appropriate method.

  FIG. 3 schematically shows a coordinate system with reference to north. The x-axis points north N. The z axis extends parallel to the gravity axis g. The twist about the x-axis is measured by the roll angle Φ (cant angle). A twist about the z-axis corresponds to an azimuth angle Ψ (also referred to as a heading angle, a heading angle or a yaw angle), and a twist about the y-axis is a tilt angle (pitch angle or pitch cone angle) Θ. provide. The three differential angles ΔΦ, ΔΨ and ΔΘ that generate the vehicle direction with respect to the coordinate system relative to the north are measured by the inertial navigation device INS. The unit vector in the machine direction is

As obtained.

The measured values of the inertial navigation device INS are read at equal intervals over the length of the curve to be measured (generally 0.25 to 1 m). At this time, the first data point P _i is simply calculated as follows.

At the next measurement, a new point P _{i + 1} is generated by vector addition.

In this way, all data points to be measured are generated. The mileage (curve length s) is measured and recorded by an odometer,
si = i · λ
i is obtained as the number of data points to be read.

  Since the equidistant measurement interval λ is very short, errors caused by secant lines can be ignored. The rail can be considered straight within a typical measurement interval of 0.25 to 1 m. The wavelength of the error to be considered is generally about 100 m or less. Therefore, in some cases, the combined error can also be ignored.

FIG. 4 schematically shows a three-dimensional coordinate line 3 having the coordinate values P _i (x _i , y _i , z _i ) as described above in a coordinate system based on the north. By setting z to 0, this three-dimensional coordinate line is projected onto the xy plane, that is, the so-called position graph 1. By setting x to 0, this three-dimensional coordinate line is projected onto the yz plane, i.e. the gradient graph. The position graph 1 shows a trajectory transition, a straight section, a transition curve, or a total arc length. The trajectory direction error is superimposed on this curve (Geometry). The inclination graph in the vertical axis direction shows the absolute inclination, the gradient change point, and the superimposed height error. The actual curve length s _i (trajectory km) is stored together for each coordinate value P _i measured and recorded as a parameter.

FIG. 5 schematically shows a position graph (z = 0). In order to draw the target position curve _RS , according to the present invention, for example, the actual position curve 3 is divided into a plurality of equally spaced sections D. In the first step, for example, an interpolated three-dimensional regression spline curve or higher-order mathematical regression curves SPL1, SPL2, etc. (according to the least square method) are calculated for each section D. In the following, the procedure for a three-dimensional regression spline curve is exemplarily described. For example, a cubic three-dimensional regression spline curve is used so that an inflection point (for example, an inflection point from the left curve to the right curve) can be correctly displayed in the trajectory shape. The length of the section D is selected according to the track grade. The faster you travel, the larger the radius and the smaller the trajectory error. Therefore, for higher track grades, longer sections D are also selected. In the next step, a plurality of transition lengths L _u are selected around x ₁ and x ₂ . Similarly, the lengths of these transition lengths _Lu are also selected according to the line class. Therefore, the plurality of regression curves can be combined with each other so that successive transitions occur under boundary conditions. These binding curves SPL3 can be created by higher order mathematical interpolation curves. Here, the three-dimensional interpolation polynomial SPL3 will be described as an example. However, other mathematical equations are possible. A plurality of points on the interpolation spline curve the calculated (P2 due to P1 and the spline equations due to the spline equations SPL 1 SPL 2) is calculated by _{x 1} or _{x 2.} In this case, as shown enlarged in FIG. 6, the 3-dimensional polynomial SPL3, it is calculated between the point P ₁ and point P ₂ under initial condition. For a single three-dimensional transition polynomial, the slopes t ₁ and t ₂ at these points of the points P 1 and P 2 and the three-dimensional regression spline curve are known. A continuous and smooth target position graph R _S drawn mathematically uniquely is obtained by this step. This target position graph _RS is second-order differentiated, and at this time, the curvature k = 1 / R of the target position as used in the railway is obtained. This has significant advantages. This is because a target shape that can be newly used in the next track maintenance work is calculated. FIG. 6 also schematically illustrates a curvature graph k (s) with respect to the target shape curve y (x). The following equation is used as a three-dimensional polynomial.
y = a · x ³ + b · x ² + c · x + d

For the tangent of the end point,
t = 3 · a · x ² + 2 · b · x + c
Is used.

P1 = (x ₁ , y ₁ ); P2 = (x ₂ , y ₂ ) is used for the plurality of points.

  At this time, the parameters a, b, c, and d can be analyzed and calculated as follows.

  The least squares method for regression with a three-dimensional spline curve uses the following four normal equations.

In order to calculate the coefficients a ₀ , a ₂ , a ₃ and a ₄ , the system of equations needs to be solved, for example by Gaussian algorithm or according to Kramel's rule.

In this case, the three-dimensional spline curve obtained according to the least square method is
y = a ₀ + a ₁ · x + a ₂ · x ² + a ₃ · x ³
In the schematic diagram, the position of the string is also entered by the length L of the string and the string sections a and b.

Figure 7 shows how a target position curve R _S may be adapted relative to a fixed point Z _P schematically. First, as described above, the target position curve R _S is calculated from a plurality of points of the actual position curve R _i . The interval d _P is retracted or advanced from the fixed point and overlaps the transition spline curve with boundary conditions P ₁ , t ₁ and Z _P , t _{2 ′} . In this case, the tangent t _{2 ′} is selected parallel to the tangent t ₂ . The curve portion after that is calculated in the same manner.

Accordingly, the target position curve, to always pass through the fixed point Z _P, the target position curve is changed. At this time, the displacement v is calculated as R _S −R _i .

FIG. 8 schematically shows a procedure for fitting the target position curve _RS when the fixed position W needs to be observed. This example demonstrates the (length _{L W)} Tentetsuki. In the switch, the target position curve must extend from (W _A ) to (W _E ) along the tangent line. Therefore, already P _A and P _E point calculated target position curve is provided. Here, as described above, a three-dimensional spline curve is newly calculated, and the first tangent (t _A , t _W ), the last tangent (t _E , t _W ), and the start point (P _A ) of the three-dimensional spline curve are calculated. , W _E ) and end point (W _A , P _E ) are preset. L _kA and L _kE are selected so that the deviation between the new target position curve (R _{S ′} ) and the preceding target position curve (R _S ) does not always significantly affect the running of the vehicle. At this time, the string of the maintenance machine moves numerically over the obtained final target position curve (R _{S ′} ). As a result, the transition of the target deviation is obtained. The deviation between the target position curve (R _{S ′} ) and the actual position curve (R _i ) provides a correction value for the direction.

  FIG. 9 schematically shows an actual height graph (x = 0) 15. The gradient change point NW can be recognized in the transition of height by a sharp curve. For example, a regression function (for example, a linear regression line) is calculated between a plurality of gradient change points (a constant uphill or downhill). Other regression functions can also be used. In the following, the method is illustrated based on a linear regression line. In order to more accurately determine the position of the gradient change point, the height graph 15 is first filtered by a digital filter so that short wavelength components (eg up to a wavelength of 50 m) are filtered. A smooth height transition 17 is obtained from the filtering. The filtered height transition 17 is subjected to numerical second order differentiation. A clear peak value that can be easily detected is generated at the gradient change point by the second-order differentiation. The maximum value (the center of the maximum gradient portion at the gradient change point) indicates the position of the gradient change point.

Calculated by NW _A and NW _E points present in the regression straight line is, the thus calculated has been the position of the gradient change point NW, the left and right gap formed between the radius of curvature and the curvature radius (length of generally 6 m) Is done. These points are then joined by, for example, a three-dimensional spline curve 18. Tangent and point NW _A and NW _E itself of the regression line at point NW _A and NW _E is applied as a boundary condition for the cubic spline curve 18. As a result, the target height graph H _S of the vertical axis direction is obtained. Then, track maintenance machine chord L, a, b are numerical moved along the curve in the longitudinal direction of the target height on the graph H _S. At this time, as a result, the target height f _L of the vertical axis direction is obtained. The deviation between the target height graph in the vertical axis and the actual height graph in the vertical axis provides a correction value for the height Δz.

The lower graph shows how the raised value (Hebewerte) is calculated from the deviation Δz. The maintenance machine performs only Hebgen 20 and does not perform Setzungen (the negative sign of the correction value). Therefore, the negative maximum correction value | H | is calculated within the range of the linear regression line. This value is added to all the calculated correction values. A new reference line N _L is obtained. Thus, all correction values H can be made positive and executed by the track maintenance machine. When the calculated bottom-up correction value (25) exceeds the allowable maximum bottom-up value _Hmax , the area (hatched area) in the excess is easily limited to the maximum bottom-up value.

DESCRIPTION OF SYMBOLS 1 Wheel, position graph 2 Wheel pair, inclination graph, longitudinal section 3 Wheel pair, coordinate value, actual position curve 4 Torsion shaft 5 Inertial navigation device 6 Lateral direction 7 Distance sensor 8 Measurement vehicle frame 9 Vertical cylinder 10 Inclination cylinder 11 Friction Bearing 12 Rail 14 Pneumatic cylinder 15 Actual height graph 17 Smooth height transition, filtered height transition 18 3D spline curve 20 Bottom up, correction line 21 Machine frame, track position to be corrected 24 Odometer (Odometer) )
25 Bottom raising correction value A Trajectory measurement vehicle B High precision trajectory correction mechanism D Equally spaced section N North W Fixed position NW Gradient change point SPW Gauge RAP Wheel interval AR Correction mechanism operating direction RS Optimized trajectory target position and target position curve , Target position graph RI actual position curve SPL1 spline equation SPL2 spline equation SPL3 three-dimensional polynomial Φ roll angle Θ tilt angle Ψ azimuth angle

Claims (10)

One facility (7, 6, 9) for optimizing the track position and for pressing at least two axles, one navigation device and wheels (1) supported on one machine frame (21) against the rail (12) , 10) and a odometer (24) for measuring the length (s) of the curve of the rail (12), the vehicle travels on the track after the measurement run by the track measuring vehicle (A) capable of running on the track. In a method for manipulating possible railway maintenance machines,
First, a measurement run is performed by the track measuring vehicle (A) capable of traveling on a track with an inertial navigation device (INS, 5), the wheels (1) of the track measuring vehicle (A) being The angular position of the inertial navigation device (INS, 5) with respect to the Cartesian coordinate system (x, y, z) is pressed by the rail (12), the length of the curve is measured by the odometer (24) (ΔΨ, ΔΦ, ΔΘ) is measured and stored, and
The trajectory of the inertial navigation device (INS, 5) in the coordinate space after the measurement run
Coordinate point (P _i ) is calculated from the stored angular positions (ΔΨ, ΔΦ, ΔΘ) by interpolation calculation, and the trajectory
From the coordinate point (P _i ) thus optimized, the position graph (1) is created by setting the z component to zero, and the longitudinal section (2) is obtained by setting the x component to zero. The method, characterized in that a curve generated and then calculated by a mathematical combination function is drawn over the coordinate points (P _i ). A plurality of higher order regression functions (SPL1, SPL2) are drawn over the coordinate points (P _i ), and this mathematical combination function (SPl2) is the end point (P ₁ , SPL2) of these regression functions (SPL1, SPL2). P ₂₎ and these regression function (SPL 1, the end nodes of _{SPL2) (t 1, t 2} ) and so as to extend through these regression function adjacent to each other within one range _{(L u)} Are connected to one higher-order mathematical combination function (SPL3), and then the fixed point (Z _P ) determines the boundary condition, one target position curve is the angle correction value (Δφ) adapted by superimposing a starting point _{(W a)} and end point (W, _{L W)} fixed position to determine the boundary conditions for the (W) is higher order plurality of mathematical combination function (22, 23) One fiber length (L _kA , L The method of claim 1, wherein the method is adapted over _kE ). The method according to claim 1 or 2, characterized in that the tolerance is preset for a fixed position (W) and a fixed point (Z _P ) that determine the boundary conditions. A plurality of gradient change points (NW) are calculated in the longitudinal cross-sectional view (2), ie, the height graph (15), and a plurality of regression functions (H _Si ) are calculated between the plurality of gradient change points. Being calculated for the interval,
For one target height curve (H _Si ), a plurality of gradient change point transition points (NW _A , NW _E ) are drawn by a plurality of mathematical transition functions, and a bottom-up correction value is defined positively. Thus, the entire correction line (20) is raised by the maximum value of “sink” (H), or the absolute value (| H |) becomes the maximum value of “sink” (H). It is added, the method according to any one of claims 1 to 3, characterized in that no maximum permissible raised value (H _max) greater than the value (25) is cut out. One working string (a, b, P _H , P _a , P _V ) of the track maintenance machine is calculated and moved over one target position curve and one target height transition, and a target deviation (f _S ) and the target height (f _L , P _Lh , P _Lv , L, a, b) in the vertical axis direction are calculated. At this time, the target position curve (R _S ) and the actual position curve (R _i ) are calculated. ) To generate a correction value for the direction (v), and the difference between the target height transition and the actual height transition to generate a correction value (H _L ) for the bottom elevation, thereby triangulation 5. Control signals (v, H _L and f _S and f _L ) necessary for controlling the track maintenance machine according to the above are transmitted to the control system of the track maintenance machine. The method according to item. The analysis mathematical graph of one target position curve (SPL1, SPL2, SPL3) is differentiated twice, and the curvature (1 / R) obtained from the two differentiations is the normal curvature graph (k ₁ , K ₂ ) depending on the length of the curve, the target height (H _S ) is similarly differentiated twice and displayed as a curvature graph of the height depending on the length of the curve. 6. The method according to any one of claims 1 to 5, wherein: An orbit measuring vehicle (A) having an inertial navigation device or an inertial measurement system (INS; 5) is towed behind a track maintenance machine, and an orbital position error (H _L , v), a cant error (Δu) and an orbital width error ( The method according to claim 1, wherein ΔSPW) is recorded as a residual error maintained after work during inspection recording. One facility for optimizing the track position and for pushing at least two axles, one navigation device and wheels (1) supported on one machine frame (21) into a track with two rails (12) Measurement by a track measuring vehicle (A) capable of traveling on a track having (7, 6, 9, 10) and a odometer (24) for measuring the length (s) of the curve of the rail (12). In the equipment for maneuvering the railway maintenance machine that can travel on the track after traveling,
The track measuring vehicle (A) comprises a plurality of wheel groups each having at least two wheels (1) each assigned to one rail in the region of the opposing longitudinal sides, these wheel groups being The track measuring vehicle (A) is supported so as to be twistable with respect to one track measuring vehicle horizontal axis (4), and the track measuring vehicle (A) includes an inertial navigation device (INS; 5) and the wheel (1). A facility (7, 6, 9, 10) for pressing against the rail (12) and an odometer (24) for measuring the length (s) of the curve of the rail (12). The equipment in question.   The equipment for pressing the wheel (1) against the rail (12) has an adjustment drive unit, by which the wheel group causes the wheel rim of the wheel (1) to move to the rail head. 9. Equipment according to claim 8, characterized in that it is slidable in the direction of the trajectory measuring vehicle horizontal axis (4) in order to press against the inner surface of the vehicle.   10. Equipment according to claim 9, characterized in that a sensor (7) assigned to the trajectory measuring vehicle horizontal axis (4) is provided for measuring the trajectory width (SPW).